diff options
| -rw-r--r-- | README.g-common | 30 | ||||
| -rw-r--r-- | README.g-cran | 12 |
2 files changed, 42 insertions, 0 deletions
diff --git a/README.g-common b/README.g-common new file mode 100644 index 0000000..2cb7f69 --- /dev/null +++ b/README.g-common @@ -0,0 +1,30 @@ +==ABOUT== +g-common is a package which is used to install packages which are not in the ebuild format as described in the Package Manager Specification, using "repository drivers" which read non-ebuild repositories and translate them into PMS-style data. Usage of these drivers is transparent and users shouldn't need to deal with these, except to install them. + +==USAGE== +If you're using a package manager which has integrated support for g-common, please refer to your package manager's instructions. Package managers known to have g-common integration: +- Portage (you can use layman to add repositories) +If you're not using such a package manager, create a new local overlay, cd into it and execute: +# g-common . sync <driver name> <remote repository uri> +where: +- . is the current directory +- <driver name> is the name of the driver used to sync the repository +- <remote repository uri> is the uri to the repository you'd like to sync +Now you can generate an ebuild tree using: +# g-common . generate-tree + +==DEVELOPMENT OF G-COMMON REPOSITORY DRIVERS== +g-common drivers are executables called as the following system call: +# <driver executable> <local overlay> <action> [arguments ...] +where: +- <driver executable> is the driver's executable's name +- <local overlay> is the location where the driver can store data in a dot-prepended directory as needed +- <action> is the action g-common wishes the driver to execute +- [arguments ...] is an optional list of arguments +Requiredly supported actions are: +- sync <uri> : download new data from the remote repository +- generate-metadata : generate any PMS-style files from the previously downloaded files before ebuilds are written by g-common +- list-packages : list all packages in a <category>/<package> <version> format (one package per line) +- list-categories : lists all the names of categories, one per line +- package <category>/<package> <version> : return a key=value list of settings for some package, no quotes, to be parsed by "export" line by line +- any PMS-supported phase (src_prepare, pkg_unpack, pkg_postinst, etc..) : executes whatever needs to happen in that PMS phase. The environment variables are filled according with the package that's being built. Nonzero return code results in a die of the ebuild diff --git a/README.g-cran b/README.g-cran new file mode 100644 index 0000000..b5be64e --- /dev/null +++ b/README.g-cran @@ -0,0 +1,12 @@ +==ABOUT== +G-CRAN is a package which is used to install packages from CRAN-style (Central R Archive Network) repositories, that is, from R package repositories. R, in turn, is a language used for statistical and biocomputational research and development. Hence, there are also two repositories alongside CRAN called the Bioconductor repositories, which contain biological code and data. + +==USAGE== +To use G-CRAN, you need to use g-common, or, if your package manager has integrated g-common support, whatever system your package manager uses. Simply install G-CRAN (note that G-CRAN depends on R) and g-common, and use g-common as described in its documentation. + +==EXAMPLE REPOSITORIES== +Some supported repositories include: +- CRAN: http://cran.r-project.org/ +- Bioconductor code: http://bioconductor.org/packages/devel/bioc +- Bioconductor annotation: http://bioconductor.org/packages/devel/data/annotation +Please note that the Bioconductor annotation repository depends on the code repository, so please install both if you'd like to use the Bioconductor databases. |